Participants

People with OA knee enquiring consecutively to Arthritis WA about access to appropriate services were invited to participate in the new disease specific self-management program. Those who were not interested in the OAK program, did not meet the selection criteria, or were not confident they would be able to participate fully were encouraged to utilize other appropriate services of Arthritis WA.

Only those with a diagnosis of OA of the knee were enrolled. It was a requirement that the diagnosis was confirmed by the participant's medical practitioner. Diagnostic criteria were at the discretion of the doctor. Disease severity was not a selection criterion. Unilateral total knee replacement did not preclude enrolment. Other criteria for ineligibility was age greater than 85 years; inability to walk 300 meters; inflammatory joint disease including rheumatoid arthritis; major concurrent illness such as cancer; bilateral knee replacement; knee surgery scheduled within 6 months of commencing the program; or physical impairments that precluded fulfilling the requirements of the program. Those people were referred to other available services.

This project was consistent with the National Health and Medical Research of Australia definition for a quality assurance project [11]. The OAK program and the associated clinical assessments were clearly described to all volunteers who had the opportunity to have all questions answered and provided verbal consent to participate.

Intervention

Groups of 8–10 participants attended 6 education sessions (one 2.5-hour session per week). Participants were provided with written material relevant to the information component discussed each week. The program used a holistic approach, including a range of aspects of care such as:

• Pain management strategies

• Exercise advice

• Joint protection

• Medication/analgesia

• Balance and falls prevention

• Coping with negative emotions

• Fatigue

• Self-management skills (goal setting, problem solving, cognitive techniques)

The fidelity of the OAK program was maintained by the use of a facilitator's manual with modules for program delivery each week. The program was delivered by 2 nurses and assessments were performed by 3 physiotherapists who had no contact with the participants other than during the assessment sessions. Participants were assessed by the same physiotherapist whenever possible to ensure consistency. The assessors did not participate in the facilitation of the program. It was a requirement that health professionals who delivered the program meet minimum musculoskeletal knowledge requirements.

Attendance was recorded at each of the 6 intervention sessions and at each assessment time-point.

Response to Intervention

Participants were assessed at baseline, immediately post-intervention at 8 weeks, and at 6 and 12 months after the program. In addition, pain was assessed on a week-to-week basis during the first 8 weeks using a VAS.

Assessments included

Health status was assessed using both a disease specific and a generic index as follows:

• The disease specific WOMAC Osteoarthritis Index (WOMAC LK3.0) assesses pain, stiffness and physical function in people with OA of the hip or knee [12]. Validity and reliability of the WOMAC pain, physical function and stiffness subscales are well established and the questionnaire is sufficiently sensitive to detect changes in health status in response to intervention [12].

• The generic Medical Outcomes Study Short Form 36 Version 1 questionnaire (SF-36). The SF-36 was designed to provide a profile of scores that would be useful in understanding the health burden in chronic diseases and the effect of treatment on general health status. It includes 8 component sub-scales that correspond to aspects of physical and mental health and well being [13]. Adequate reliability for between group comparisons has been demonstrated in numerous studies and an English version of the questionnaire has been developed and validated specifically for use in Australia [14]. For people with OA, an improvement of 5 points on the physical component score of the SF-36 is considered to be clinically significant [15].

Pain was assessed using pain scales included in the WOMAC and SF-36 indices. During the intervention period, pain was monitored on a week-to-week basis (Figure 1) using a 100 mm VAS. The left hand anchor was "No Pain", and the right hand anchor "Worst Pain". The VAS is well established in clinical practice for measuring pain levels post-surgery, following drug therapy and other interventions in arthritis populations [16]. A reduction of 30% or 2 points in VAS is considered to represent a clinically important difference [17, 18].

Active range of motion of knee flexion and extension were measured using a long armed goniometer [19]. The reliability and validity of the goniometer is well established for measurement of active knee flexion and extension [19, 20].

Balance was assessed using a timed single leg balance test. This simple test assesses the length of time, to a maximum of 10 seconds, a person can stand on one leg. It is a good predictor of falls in the elderly [21] and is reliable and valid [22].

Statistical Analysis

Data were analysed using SPSS v16 for Macintosh. One-way (repeated measures) analysis of variance with time (baseline, 8 weeks, 6 months and 12 months) as the independent variable was used to assess changes in the variables of interest. Where the ANOVA was significant pair-wise differences between baseline and 12 months were compared using paired t-tests and mean change and 95% confidence interval for were calculated. The effect size for the pair wise comparison was calculated using Cohen's d (the difference between the means: M₁ - M₂ divided by the standard deviation). Improvement is represented as a positive difference: small, d = 0.2; medium d = 0.5; large d = 0.8 [23] allowing the comparison of outcomes across the intervention. Separate models were constructed for each outcome variable. Normal distribution and homogeneity of the variance were confirmed prior to further analysis. Statistical significance was inferred at a 2-tailed p < 0.05.

Pain

In this study pain was measured in a number of ways, all demonstrating an improvement. A number of aspects of the self-management intervention may have contributed to the reduced pain levels reported by participants. Both aerobic and resistance exercise in a home-based exercise program have been shown to significantly reduce knee pain in-patients with OA [29, 30]. An important component of the OAK intervention is discussion on the formulation of a comprehensive home exercise program that incorporates strengthening, endurance, balance and flexibility components. Participants were not taken into a gym or given individual personal training; however they were encouraged to pursue that option independently.

The exercise component was not controlled and participants freely chose the type of exercise/s and the degree to which they would comply. By providing a number of exercise alternatives, it was hoped that exercise routines would become habitual by the end of the 6-week program. In accordance with self-management principles, participants were motivated to use their "library" of exercise choices when planning their weekly goals. The use of goal setting with participants promoted good adherence to the exercise program, as reported each week, but data regarding adherence were not collected for this study.

As well as exercise instruction and cognitive therapies, medication usage and therapeutic dosing principles in particular for analgesia were taught to encourage medication compliance and effective pain management. The average age of participants was 66 years and most had several co-morbidities requiring medication (Table 1). Many participants had an aversion to medications and delayed taking analgesia until their pain was acute and therefore more difficult to control. Pain management guidelines were discussed with the aim of determining patterns of pain. For example short term "around the clock" analgesia dosing for acute pain, or "as needed" analgesia for intermittent pain.

It is likely that the OAK intervention has facilitated better pain-coping skills that are important predictors of disability associated with OA. Previous studies have reported that catastrophizing and negative self-statements are associated with increased knee pain [31]. In the OAK intervention, participants were taught strategies for cognitive symptom management such as distraction, guided imagery, relaxation and thought challenging techniques that are considered to be important additional measures of pain management in people with OA [30, 32].

Health Status

Participants reported considerable improvements in physical function. Like pain, functional improvements were reflected by changes in a number of the parameters measured. It is generally accepted that the WOMAC questionnaire has greater specificity and consequently better responsiveness for people with OA [33], nonetheless, the SF-36 also reflected these changes.

Interpreting these results requires some understanding of the value patients place on improvements of this magnitude. Establishing this can be difficult. A number of methods, each with strengths and limitations, have been used but findings are not entirely consistent. Improvements of 9% to 10% in WOMAC scores in response to rofecoxib or ibuprofen were perceptible to patients with OA knee [34] when anchored against a patient global assessment of response to therapy. Changes observed in our study were generally more than twice this magnitude. On the other hand the 21.6% improvement in WOMAC function was somewhat less than 26%, the minimal level suggested by Tubach et al [35] as clinically important.

Expressed as effect sizes in standard deviation units the improvements in the WOMAC pain and SF-36 bodily pain domains would be considered moderate [23]. The consistency of this effect between different outcome tools supports the validity of the change. Effect sizes for the WOMAC functional domain and for the SF-36 mental health domains were slightly lower at 0.4. Notably these effect sizes are larger or comparable to the pooled effect sizes for general pain from systematic reviews of NSAID therapy [36] and aerobic walking [37] (0.33 and 0.52 respectively), although larger effects are often observed in uncontrolled studies. Further context for interpretation of the improvements we observed in quality of life measured by the SF-36 may be provided by considering the average decline of 2.1 points over 12 months reported in people with OA in this age group [15].

Limitations

The subjects who attended this quality assurance program were typical of those who attend self-management programs run by Arthritis WA. Over representation in the highest socio-economic group (Table 1) may affect the reproducibility of this program, however, the demographics of the area this program was conducted in are consistent with this attendance statistic. These results should be interpreted with caution as this limits the generalization to other socio-economic groups. Testing the OAK program with other socio-economic groups was outside the limitations of this QA program.

It is important to note that no control period or control group were available for comparison. Consequently, the clinical improvements observed in this cohort should be interpreted with caution. Despite this, improvements in response to this disease specific self-management program delivered by health professionals are encouraging and have interest. We therefore propose to further evaluate the benefits of this program using a more rigorous study design.